docs: fill in discovered platform env values and add .env.example

[haasonsaas]

Replace example URLs and placeholders with actual values discovered from evalops/deploy and evalops/platform:

• Gateway URL: https://llm-gateway.evalops.dev/v1 (live, auth-gated)
• Platform URL: https://api.evalops.dev
• Organization ID: org_evalops
• Workspace ID: evalops

Also:

• Add .env.example for quick setup (with explanation of which vars are live vs. EVA-157-gated)
• Add .env to .gitignore
• Document that registration and span ingest are blocked on EVA-157 — agent-mcp is not publicly exposed yet (all feature flags at 0%)
• Clarify that current registration/span-ingest defaults are in-process MCP stubs (mcp / mcp:self_diagnostic)

Gateway works today with a valid identity token. Registration + span ingest come online when EVA-157 ships.

[cursor]

PR Summary

Low Risk
Documentation and sample env only; no runtime or security logic changes.

Overview
Adds .env.example with real EvalOps URLs (llm-gateway.evalops.dev, api.evalops.dev), org/workspace IDs, agent identity fields, and commented safety/endpoint knobs so operators can copy to .env and only fill tokens.

.gitignore now ignores .env so local secrets are not committed.

docs/install-config.md replaces placeholder example URLs with the discovered production values, notes that platform registration/span ingest are EVA-157-gated, and documents that default registration/span ingest use in-process MCP stubs (mcp / mcp:self_diagnostic) versus the post-EVA-157 span ingest path.

^{Reviewed by Cursor Bugbot for commit 34614c3. Bugbot is set up for automated code reviews on this repo. Configure here.}